This is the graded finale for the whole course. It pulls every thread together — the core silent-evidence mechanic, Abraham Wald and the returning bombers, the model hiding in mutual-fund tables and “habits of the successful” listicles and old buildings and falling cats, the cognitive roots that make the trap feel invisible, and the denominator-hunting defences that break it. Read each question twice: several look easy until you notice that the sample has been quietly filtered, and the tempting answer is almost always the one that trusts the survivors.
How this exam works
Questions appear one at a time. Once you submit an answer it LOCKS — there is no going back, no retry, and no restart, so commit before you click. Your score stays hidden until the very end, when you find out whether you cleared the bar. The pass mark is 70 percent. A few questions ask you to select ALL correct options, not just the single best one — read the prompt carefully.
In one sentence, what is the core mechanic of survivorship bias?
Select an answer to continue.
Course Recap
Big picture
Survivorship Bias — the whole course on one map
- Survivorship Bias
- 01 · The Silent-Evidence Mechanic
- The core idea
- A selection filter deletes part of the sample before you observe it, so the survivors you see are not a fair representation of everything that started. The failures are not hidden by lying — they are physically absent, generating no alarm. This "silent evidence" is the whole bias: the mistake is reading the filtered survivors as if they were the full population.
- The core idea
- 02 · Wald's Bombers
- Armour the gaps, not the holes
- Returning WWII bombers showed holes on wings and fuselage but clean engines and cockpits. Wald saw the clean spots were fatal hits — those planes never came back — so he armoured the low-hole regions. Survivor holes are roughly hit-rate times (1 minus lethality): a deadly, often-hit area looks deceptively clean on survivors. The survivor map is the inverse of the danger map.
- Armour the gaps, not the holes
- 03 · Survivorship in the Wild
- The model hiding everywhere
- Dead mutual funds get dropped from databases, inflating reported returns by roughly 1 to 1.5 points per year. "Habits of successful people" and business bestsellers study only winners and never the failures who shared the same traits. "Old buildings were built better" ignores the flimsy ones that collapsed. High-rise cat data looks survivable because cats killed on impact were never brought in (debated). War and Titanic testimony comes only from those alive to tell it.
- The model hiding everywhere
- 04 · Why We Fall For It
- The cognitive roots
- WYSIATI — what you see is all there is — means the mind builds confident stories and raises no alarm for absent data. The availability heuristic makes vivid, publicised success easy to recall while quiet failures vanish. We hunger for clean causal stories, and winners self-select into the platforms and interviews, filtering whose voice we hear. None of these instincts naturally demand the denominator.
- The cognitive roots
- 05 · The Defences
- Find the denominator
- Ask the master question: what happened to EVERYONE who started, not just the ones who made it? Find the denominator (winners over all who tried), hunt the missing cohort, and reconstruct the filter — name the mechanism (closure, death, dropout, delisting) that removed cases, then add them back. Use control groups and prefer full-population data over survivor testimonials. Collecting more winners never fixes it; you must recover the failures.
- Find the denominator
- 01 · The Silent-Evidence Mechanic
Key takeaways
Survivorship bias is not lying and not bad luck — it is a filter that quietly deletes the failures before you look, leaving a gallery of survivors you then mistake for the whole population. That one mechanic explains Wald armouring the bullet-free engines, mutual-fund tables inflated by their vanished dead, “successful habits” listicles blind to the millions who did the same and failed, sturdy old buildings that are merely the ones that lasted, cats that “survive” high falls because the fatal ones never reached the vet, and survivor testimony from wars and shipwrecks. The reason it feels invisible is that the mind raises no alarm for absent data (WYSIATI), the vivid winners are the easiest to recall (availability), and the platform always goes to those still standing. Your defences are always the same shape: find the denominator, hunt the missing cohort by asking “what happened to everyone who started?”, reconstruct the filter that removed them, insist on a control group, and prefer full-population data over the testimony of the survivors. Piling up more winners never cures it — only recovering the failures does.